Previous research in this laboratory and others has supported the prevailing noesis that the tumor neovasculature is the primary target tissue for the anti-tumor effects of TNF. The selective effects of TNF on tumor neovasculature during clinical isolation perfusion are striking; within hours after treatment there is eschar formation occuring in superficial tumors and rapid liquefaction necrosis of deep-seated ones. We have shown that retroviral transduction and over-expression of a tumor derived cytokine, endothelial monocyte activating polypeptide (EMAP) II, will alter the in vivo phenotype of a TNF insensitive melanoma to a TNF sensitive one without altering its sensitivity to TNF cytotoxicity in vitro. The mechanism for this phenomenon is likely due to the sensitization of tumor neovasculature to the procoagulant effects of TNF. This observation along with the growing understanding of a dynamic host to tumor interaction led us to hypothesize that within the tumor microenvironment there are proteins produced by infiltrating cells, resident fibroblasts, endothelium, and tumor cells which results in a local milieu that influences tumor progression, invasion, and metastases. We have identified 2 candidate endogenous proteins that may have indirect anti-tumor properties by effects within the tumor microenvironment, interferon gamma inducible protein 10 (IP-10) and the interleukin 1 receptor antagonist (IL-1ra). The specific research aims of this project are: 1) Characterize the anti-tumor properties of IP-10. 2) Characterize the anti-tumor properties of IL-1ra. Characterize the anti-tumor properties of IP-10. IP-10 is a C-X-C chemokine that is produced by activated monocytes, fibroblasts, endothelial cells, and keratinocytes in response to stimulation with interferon-gamma and is primarily a chemotactic factor for T-lymphocytes. It has in vivo anti-tumor effects in some experimental models that have been proposed to be mediated via a T-cell dependent mechanism. We evaluated the effects of retroviral IP-10 gene transduction and protein over-expression in human A375 melanoma xenografts in nude mice using a pCLNCX retroviral vector containing a CMV promoter to drive the IP-10 transgene with Null transfected cells (Null) and Wild Type (WT) tumor lines as controls. Two transfectants with high IP-10 protein expression were identified and used in subsequent experiments. There was no difference on in vitro proliferation rates of the various clones compared to Null transfectants or WT cell lines. However, after subcutaneous injection of WT A-375 cells, Null transduced or IP-10 transduced clones, there was significant inhibition of IP-10 tumor growth compared to control or parental tumors. Mixing experiments demonstrated a paracrine bystander effect from IP-10 produced in transduced cells; tumors consisting of 50 percent IP-10 transduced cells were significantly smaller than tumors consisting of 50 percent Null transduced cells. These data suggested that systemic administration of the recombinant protein may have similar anti-tumor properties. Microvessel counts in IP-10 transduced clones were markedly less than those in control tumors based on vonWillebrand factor immunostaining. Additional studies were undertaken to characterize the alterations in gene expression profiles in IP-10 transduced melanoma xenografts compared to Null transfectants and WT xenografts using cDNA microarray analysis. Genes that were differentially expressed by at least 3 fold between IP-10 transduced cells and both WT and Null transfectants with a less than 1.5 fold difference between WT type and Null transfectants were considered significant.